[docs] Add safetensors flag (#4245)

* add safetensors flag

* apply review

docs/source/en/optimization/fp16.md

@@ -51,6 +51,7 @@ from diffusers import DiffusionPipeline
 pipe = DiffusionPipeline.from_pretrained(
     "runwayml/stable-diffusion-v1-5",
     torch_dtype=torch.float16,
+    use_safetensors=True,
 )
 pipe = pipe.to("cuda")
 
@@ -80,6 +81,7 @@ from diffusers import StableDiffusionPipeline
 pipe = StableDiffusionPipeline.from_pretrained(
     "runwayml/stable-diffusion-v1-5",
     torch_dtype=torch.float16,
+    use_safetensors=True,
 )
 pipe = pipe.to("cuda")
 
@@ -106,6 +108,7 @@ from diffusers import StableDiffusionPipeline, UniPCMultistepScheduler
 pipe = StableDiffusionPipeline.from_pretrained(
     "runwayml/stable-diffusion-v1-5",
     torch_dtype=torch.float16,
+    use_safetensors=True,
 )
 pipe.scheduler = UniPCMultistepScheduler.from_config(pipe.scheduler.config)
 pipe = pipe.to("cuda")
@@ -133,6 +136,7 @@ from diffusers import StableDiffusionPipeline
 pipe = StableDiffusionPipeline.from_pretrained(
     "runwayml/stable-diffusion-v1-5",
     torch_dtype=torch.float16,
+    use_safetensors=True,
 )
 
 prompt = "a photo of an astronaut riding a horse on mars"
@@ -157,6 +161,7 @@ from diffusers import StableDiffusionPipeline
 pipe = StableDiffusionPipeline.from_pretrained(
     "runwayml/stable-diffusion-v1-5",
     torch_dtype=torch.float16,
+    use_safetensors=True,
 )
 
 prompt = "a photo of an astronaut riding a horse on mars"
@@ -189,6 +194,7 @@ from diffusers import StableDiffusionPipeline
 pipe = StableDiffusionPipeline.from_pretrained(
     "runwayml/stable-diffusion-v1-5",  
     torch_dtype=torch.float16,
+    use_safetensors=True,
 )
 
 prompt = "a photo of an astronaut riding a horse on mars"
@@ -205,6 +211,7 @@ from diffusers import StableDiffusionPipeline
 pipe = StableDiffusionPipeline.from_pretrained(
     "runwayml/stable-diffusion-v1-5",
     torch_dtype=torch.float16,
+    use_safetensors=True,
 )
 
 prompt = "a photo of an astronaut riding a horse on mars"
@@ -267,6 +274,7 @@ def generate_inputs():
 pipe = StableDiffusionPipeline.from_pretrained(
     "runwayml/stable-diffusion-v1-5",
     torch_dtype=torch.float16,
+    use_safetensors=True,
 ).to("cuda")
 unet = pipe.unet
 unet.eval()
@@ -330,6 +338,7 @@ class UNet2DConditionOutput:
 pipe = StableDiffusionPipeline.from_pretrained(
     "runwayml/stable-diffusion-v1-5",
     torch_dtype=torch.float16,
+    use_safetensors=True,
 ).to("cuda")
 
 # use jitted unet
@@ -389,6 +398,7 @@ import torch
 pipe = DiffusionPipeline.from_pretrained(
     "runwayml/stable-diffusion-v1-5",
     torch_dtype=torch.float16,
+    use_safetensors=True,
 ).to("cuda")
 
 pipe.enable_xformers_memory_efficient_attention()

docs/source/en/optimization/torch2.0.md

@@ -39,7 +39,7 @@ pip install --upgrade torch diffusers
     import torch
     from diffusers import DiffusionPipeline
 
-    pipe = DiffusionPipeline.from_pretrained("runwayml/stable-diffusion-v1-5", torch_dtype=torch.float16)
+    pipe = DiffusionPipeline.from_pretrained("runwayml/stable-diffusion-v1-5", torch_dtype=torch.float16, use_safetensors=True)
     pipe = pipe.to("cuda")
 
     prompt = "a photo of an astronaut riding a horse on mars"
@@ -53,7 +53,7 @@ pip install --upgrade torch diffusers
     from diffusers import DiffusionPipeline
     + from diffusers.models.attention_processor import AttnProcessor2_0
 
-    pipe = DiffusionPipeline.from_pretrained("runwayml/stable-diffusion-v1-5", torch_dtype=torch.float16).to("cuda")
+    pipe = DiffusionPipeline.from_pretrained("runwayml/stable-diffusion-v1-5", torch_dtype=torch.float16, use_safetensors=True).to("cuda")
     + pipe.unet.set_attn_processor(AttnProcessor2_0())
 
     prompt = "a photo of an astronaut riding a horse on mars"
@@ -69,7 +69,7 @@ pip install --upgrade torch diffusers
     from diffusers import DiffusionPipeline
     from diffusers.models.attention_processor import AttnProcessor
 
-    pipe = DiffusionPipeline.from_pretrained("runwayml/stable-diffusion-v1-5", torch_dtype=torch.float16).to("cuda")
+    pipe = DiffusionPipeline.from_pretrained("runwayml/stable-diffusion-v1-5", torch_dtype=torch.float16, use_safetensors=True).to("cuda")
     pipe.unet.set_default_attn_processor()
 
     prompt = "a photo of an astronaut riding a horse on mars"
@@ -107,7 +107,7 @@ path = "runwayml/stable-diffusion-v1-5"
 
 run_compile = True  # Set True / False
 
-pipe = DiffusionPipeline.from_pretrained(path, torch_dtype=torch.float16)
+pipe = DiffusionPipeline.from_pretrained(path, torch_dtype=torch.float16, use_safetensors=True)
 pipe = pipe.to("cuda")
 pipe.unet.to(memory_format=torch.channels_last)
 
@@ -140,7 +140,7 @@ path = "runwayml/stable-diffusion-v1-5"
 
 run_compile = True  # Set True / False
 
-pipe = StableDiffusionImg2ImgPipeline.from_pretrained(path, torch_dtype=torch.float16)
+pipe = StableDiffusionImg2ImgPipeline.from_pretrained(path, torch_dtype=torch.float16, use_safetensors=True)
 pipe = pipe.to("cuda")
 pipe.unet.to(memory_format=torch.channels_last)
 
@@ -180,7 +180,7 @@ path = "runwayml/stable-diffusion-inpainting"
 
 run_compile = True  # Set True / False
 
-pipe = StableDiffusionInpaintPipeline.from_pretrained(path, torch_dtype=torch.float16)
+pipe = StableDiffusionInpaintPipeline.from_pretrained(path, torch_dtype=torch.float16, use_safetensors=True)
 pipe = pipe.to("cuda")
 pipe.unet.to(memory_format=torch.channels_last)
 
@@ -212,9 +212,9 @@ init_image = init_image.resize((512, 512))
 path = "runwayml/stable-diffusion-v1-5"
 
 run_compile = True  # Set True / False
-controlnet = ControlNetModel.from_pretrained("lllyasviel/sd-controlnet-canny", torch_dtype=torch.float16)
+controlnet = ControlNetModel.from_pretrained("lllyasviel/sd-controlnet-canny", torch_dtype=torch.float16, use_safetensors=True)
 pipe = StableDiffusionControlNetPipeline.from_pretrained(
-    path, controlnet=controlnet, torch_dtype=torch.float16
+    path, controlnet=controlnet, torch_dtype=torch.float16, use_safetensors=True
 )
 
 pipe = pipe.to("cuda")
@@ -240,11 +240,11 @@ import torch
 
 run_compile = True  # Set True / False
 
-pipe = DiffusionPipeline.from_pretrained("DeepFloyd/IF-I-M-v1.0", variant="fp16", text_encoder=None, torch_dtype=torch.float16)
+pipe = DiffusionPipeline.from_pretrained("DeepFloyd/IF-I-M-v1.0", variant="fp16", text_encoder=None, torch_dtype=torch.float16, use_safetensors=True)
 pipe.to("cuda")
-pipe_2 = DiffusionPipeline.from_pretrained("DeepFloyd/IF-II-M-v1.0", variant="fp16", text_encoder=None, torch_dtype=torch.float16)
+pipe_2 = DiffusionPipeline.from_pretrained("DeepFloyd/IF-II-M-v1.0", variant="fp16", text_encoder=None, torch_dtype=torch.float16, use_safetensors=True)
 pipe_2.to("cuda")
-pipe_3 = DiffusionPipeline.from_pretrained("stabilityai/stable-diffusion-x4-upscaler", torch_dtype=torch.float16)
+pipe_3 = DiffusionPipeline.from_pretrained("stabilityai/stable-diffusion-x4-upscaler", torch_dtype=torch.float16, use_safetensors=True)
 pipe_3.to("cuda")
 
 

docs/source/en/quicktour.md

@@ -67,7 +67,7 @@ Load the model with the [`~DiffusionPipeline.from_pretrained`] method:
 ```python
 >>> from diffusers import DiffusionPipeline
 
->>> pipeline = DiffusionPipeline.from_pretrained("runwayml/stable-diffusion-v1-5")
+>>> pipeline = DiffusionPipeline.from_pretrained("runwayml/stable-diffusion-v1-5", use_safetensors=True)
 ```
 
 The [`DiffusionPipeline`] downloads and caches all modeling, tokenization, and scheduling components. You'll see that the Stable Diffusion pipeline is composed of the [`UNet2DConditionModel`] and [`PNDMScheduler`] among other things:
@@ -130,7 +130,7 @@ You can also use the pipeline locally. The only difference is you need to downlo
 Then load the saved weights into the pipeline:
 
 ```python
->>> pipeline = DiffusionPipeline.from_pretrained("./stable-diffusion-v1-5")
+>>> pipeline = DiffusionPipeline.from_pretrained("./stable-diffusion-v1-5", use_safetensors=True)
 ```
 
 Now you can run the pipeline as you would in the section above.
@@ -142,7 +142,7 @@ Different schedulers come with different denoising speeds and quality trade-offs
 ```py
 >>> from diffusers import EulerDiscreteScheduler
 
->>> pipeline = DiffusionPipeline.from_pretrained("runwayml/stable-diffusion-v1-5")
+>>> pipeline = DiffusionPipeline.from_pretrained("runwayml/stable-diffusion-v1-5", use_safetensors=True)
 >>> pipeline.scheduler = EulerDiscreteScheduler.from_config(pipeline.scheduler.config)
 ```
 
@@ -160,7 +160,7 @@ Models are initiated with the [`~ModelMixin.from_pretrained`] method which also
 >>> from diffusers import UNet2DModel
 
 >>> repo_id = "google/ddpm-cat-256"
->>> model = UNet2DModel.from_pretrained(repo_id)
+>>> model = UNet2DModel.from_pretrained(repo_id, use_safetensors=True)
 ```
 
 To access the model parameters, call `model.config`:

docs/source/en/stable_diffusion.md

@@ -26,7 +26,7 @@ Begin by loading the [`runwayml/stable-diffusion-v1-5`](https://huggingface.co/r
 from diffusers import DiffusionPipeline
 
 model_id = "runwayml/stable-diffusion-v1-5"
-pipeline = DiffusionPipeline.from_pretrained(model_id)
+pipeline = DiffusionPipeline.from_pretrained(model_id, use_safetensors=True)
 ```
 
 The example prompt you'll use is a portrait of an old warrior chief, but feel free to use your own prompt:
@@ -75,7 +75,7 @@ Let's start by loading the model in `float16` and generate an image:
 ```python
 import torch
 
-pipeline = DiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.float16)
+pipeline = DiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.float16, use_safetensors=True)
 pipeline = pipeline.to("cuda")
 generator = torch.Generator("cuda").manual_seed(0)
 image = pipeline(prompt, generator=generator).images[0]

docs/source/en/training/adapt_a_model.md

@@ -11,7 +11,7 @@ A [`UNet2DConditionModel`] by default accepts 4 channels in the [input sample](h
 ```py
 from diffusers import StableDiffusionPipeline
 
-pipeline = StableDiffusionPipeline.from_pretrained("runwayml/stable-diffusion-v1-5")
+pipeline = StableDiffusionPipeline.from_pretrained("runwayml/stable-diffusion-v1-5", use_safetensors=True)
 pipeline.unet.config["in_channels"]
 4
 ```
@@ -21,7 +21,7 @@ Inpainting requires 9 channels in the input sample. You can check this value in
 ```py
 from diffusers import StableDiffusionPipeline
 
-pipeline = StableDiffusionPipeline.from_pretrained("runwayml/stable-diffusion-inpainting")
+pipeline = StableDiffusionPipeline.from_pretrained("runwayml/stable-diffusion-inpainting", use_safetensors=True)
 pipeline.unet.config["in_channels"]
 9
 ```
@@ -35,7 +35,12 @@ from diffusers import UNet2DConditionModel
 
 model_id = "runwayml/stable-diffusion-v1-5"
 unet = UNet2DConditionModel.from_pretrained(
-    model_id, subfolder="unet", in_channels=9, low_cpu_mem_usage=False, ignore_mismatched_sizes=True
+    model_id,
+    subfolder="unet",
+    in_channels=9,
+    low_cpu_mem_usage=False,
+    ignore_mismatched_sizes=True,
+    use_safetensors=True,
 )
 ```
 

docs/source/en/training/controlnet.md

@@ -306,9 +306,9 @@ import torch
 base_model_path = "path to model"
 controlnet_path = "path to controlnet"
 
-controlnet = ControlNetModel.from_pretrained(controlnet_path, torch_dtype=torch.float16)
+controlnet = ControlNetModel.from_pretrained(controlnet_path, torch_dtype=torch.float16, use_safetensors=True)
 pipe = StableDiffusionControlNetPipeline.from_pretrained(
-    base_model_path, controlnet=controlnet, torch_dtype=torch.float16
+    base_model_path, controlnet=controlnet, torch_dtype=torch.float16, use_safetensors=True
 )
 
 # speed up diffusion process with faster scheduler and memory optimization

docs/source/en/training/custom_diffusion.md

@@ -222,7 +222,9 @@ Once you have trained a model using the above command, you can run inference usi
 import torch
 from diffusers import DiffusionPipeline
 
-pipe = DiffusionPipeline.from_pretrained("CompVis/stable-diffusion-v1-4", torch_dtype=torch.float16).to("cuda")
+pipe = DiffusionPipeline.from_pretrained(
+    "CompVis/stable-diffusion-v1-4", torch_dtype=torch.float16, use_safetensors=True
+).to("cuda")
 pipe.unet.load_attn_procs("path-to-save-model", weight_name="pytorch_custom_diffusion_weights.bin")
 pipe.load_textual_inversion("path-to-save-model", weight_name="<new1>.bin")
 
@@ -246,7 +248,7 @@ model_id = "sayakpaul/custom-diffusion-cat"
 card = RepoCard.load(model_id)
 base_model_id = card.data.to_dict()["base_model"]
 
-pipe = DiffusionPipeline.from_pretrained(base_model_id, torch_dtype=torch.float16).to("cuda")
+pipe = DiffusionPipeline.from_pretrained(base_model_id, torch_dtype=torch.float16, use_safetensors=True).to("cuda")
 pipe.unet.load_attn_procs(model_id, weight_name="pytorch_custom_diffusion_weights.bin")
 pipe.load_textual_inversion(model_id, weight_name="<new1>.bin")
 
@@ -270,7 +272,7 @@ model_id = "sayakpaul/custom-diffusion-cat-wooden-pot"
 card = RepoCard.load(model_id)
 base_model_id = card.data.to_dict()["base_model"]
 
-pipe = DiffusionPipeline.from_pretrained(base_model_id, torch_dtype=torch.float16).to("cuda")
+pipe = DiffusionPipeline.from_pretrained(base_model_id, torch_dtype=torch.float16, use_safetensors=True).to("cuda")
 pipe.unet.load_attn_procs(model_id, weight_name="pytorch_custom_diffusion_weights.bin")
 pipe.load_textual_inversion(model_id, weight_name="<new1>.bin")
 pipe.load_textual_inversion(model_id, weight_name="<new2>.bin")

docs/source/en/training/distributed_inference.md

@@ -16,7 +16,9 @@ Now use the [`~accelerate.PartialState.split_between_processes`] utility as a co
 from accelerate import PartialState
 from diffusers import DiffusionPipeline
 
-pipeline = DiffusionPipeline.from_pretrained("runwayml/stable-diffusion-v1-5", torch_dtype=torch.float16)
+pipeline = DiffusionPipeline.from_pretrained(
+    "runwayml/stable-diffusion-v1-5", torch_dtype=torch.float16, use_safetensors=True
+)
 distributed_state = PartialState()
 pipeline.to(distributed_state.device)
 
@@ -50,7 +52,9 @@ import torch.multiprocessing as mp
 
 from diffusers import DiffusionPipeline
 
-sd = DiffusionPipeline.from_pretrained("runwayml/stable-diffusion-v1-5", torch_dtype=torch.float16)
+sd = DiffusionPipeline.from_pretrained(
+    "runwayml/stable-diffusion-v1-5", torch_dtype=torch.float16, use_safetensors=True
+)
 ```
 
 You'll want to create a function to run inference; [`init_process_group`](https://pytorch.org/docs/stable/distributed.html?highlight=init_process_group#torch.distributed.init_process_group) handles creating a distributed environment with the type of backend to use, the `rank` of the current process, and the `world_size` or the number of processes participating. If you're running inference in parallel over 2 GPUs, then the `world_size` is 2.

docs/source/en/training/dreambooth.md

@@ -303,7 +303,9 @@ unet = UNet2DConditionModel.from_pretrained("/sddata/dreambooth/daruma-v2-1/chec
 # if you have trained with `--args.train_text_encoder` make sure to also load the text encoder
 text_encoder = CLIPTextModel.from_pretrained("/sddata/dreambooth/daruma-v2-1/checkpoint-100/text_encoder")
 
-pipeline = DiffusionPipeline.from_pretrained(model_id, unet=unet, text_encoder=text_encoder, dtype=torch.float16)
+pipeline = DiffusionPipeline.from_pretrained(
+    model_id, unet=unet, text_encoder=text_encoder, dtype=torch.float16, use_safetensors=True
+)
 pipeline.to("cuda")
 
 # Perform inference, or save, or push to the hub
@@ -318,7 +320,7 @@ from diffusers import DiffusionPipeline
 
 # Load the pipeline with the same arguments (model, revision) that were used for training
 model_id = "CompVis/stable-diffusion-v1-4"
-pipeline = DiffusionPipeline.from_pretrained(model_id)
+pipeline = DiffusionPipeline.from_pretrained(model_id, use_safetensors=True)
 
 accelerator = Accelerator()
 
@@ -333,6 +335,7 @@ pipeline = DiffusionPipeline.from_pretrained(
     model_id,
     unet=accelerator.unwrap_model(unet),
     text_encoder=accelerator.unwrap_model(text_encoder),
+    use_safetensors=True,
 )
 
 # Perform inference, or save, or push to the hub
@@ -488,7 +491,7 @@ from diffusers import DiffusionPipeline
 import torch
 
 model_id = "path_to_saved_model"
-pipe = DiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.float16).to("cuda")
+pipe = DiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.float16, use_safetensors=True).to("cuda")
 
 prompt = "A photo of sks dog in a bucket"
 image = pipe(prompt, num_inference_steps=50, guidance_scale=7.5).images[0]
@@ -510,7 +513,7 @@ must also update the pipeline's scheduler config.
 ```py
 from diffusers import DiffusionPipeline
 
-pipe = DiffusionPipeline.from_pretrained("DeepFloyd/IF-I-XL-v1.0")
+pipe = DiffusionPipeline.from_pretrained("DeepFloyd/IF-I-XL-v1.0", use_safetensors=True)
 
 pipe.load_lora_weights("<lora weights path>")
 

docs/source/en/training/instructpix2pix.md

@@ -165,7 +165,9 @@ import torch
 from diffusers import StableDiffusionInstructPix2PixPipeline
 
 model_id = "your_model_id"  # <- replace this
-pipe = StableDiffusionInstructPix2PixPipeline.from_pretrained(model_id, torch_dtype=torch.float16).to("cuda")
+pipe = StableDiffusionInstructPix2PixPipeline.from_pretrained(
+    model_id, torch_dtype=torch.float16, use_safetensors=True
+).to("cuda")
 generator = torch.Generator("cuda").manual_seed(0)
 
 url = "https://huggingface.co/datasets/sayakpaul/sample-datasets/resolve/main/test_pix2pix_4.png"

docs/source/en/training/lora.md

@@ -98,7 +98,7 @@ Now you can use the model for inference by loading the base model in the [`Stabl
 
 >>> model_base = "runwayml/stable-diffusion-v1-5"
 
->>> pipe = StableDiffusionPipeline.from_pretrained(model_base, torch_dtype=torch.float16)
+>>> pipe = StableDiffusionPipeline.from_pretrained(model_base, torch_dtype=torch.float16, use_safetensors=True)
 >>> pipe.scheduler = DPMSolverMultistepScheduler.from_config(pipe.scheduler.config)
 ```
 
@@ -137,7 +137,7 @@ lora_model_id = "sayakpaul/sd-model-finetuned-lora-t4"
 card = RepoCard.load(lora_model_id)
 base_model_id = card.data.to_dict()["base_model"]
 
-pipe = StableDiffusionPipeline.from_pretrained(base_model_id, torch_dtype=torch.float16)
+pipe = StableDiffusionPipeline.from_pretrained(base_model_id, torch_dtype=torch.float16, use_safetensors=True)
 ...
 ```
 
@@ -211,7 +211,7 @@ Now you can use the model for inference by loading the base model in the [`Stabl
 
 >>> model_base = "runwayml/stable-diffusion-v1-5"
 
->>> pipe = StableDiffusionPipeline.from_pretrained(model_base, torch_dtype=torch.float16)
+>>> pipe = StableDiffusionPipeline.from_pretrained(model_base, torch_dtype=torch.float16, use_safetensors=True)
 ```
 
 Load the LoRA weights from your finetuned DreamBooth model *on top of the base model weights*, and then move the pipeline to a GPU for faster inference. When you merge the LoRA weights with the frozen pretrained model weights, you can optionally adjust how much of the weights to merge with the `scale` parameter:
@@ -251,7 +251,7 @@ lora_model_id = "sayakpaul/dreambooth-text-encoder-test"
 card = RepoCard.load(lora_model_id)
 base_model_id = card.data.to_dict()["base_model"]
 
-pipe = StableDiffusionPipeline.from_pretrained(base_model_id, torch_dtype=torch.float16)
+pipe = StableDiffusionPipeline.from_pretrained(base_model_id, torch_dtype=torch.float16, use_safetensors=True)
 pipe = pipe.to("cuda")
 pipe.load_lora_weights(lora_model_id)
 image = pipe("A picture of a sks dog in a bucket", num_inference_steps=25).images[0]
@@ -307,7 +307,7 @@ import torch
 from diffusers import StableDiffusionPipeline, DPMSolverMultistepScheduler
 
 pipeline = StableDiffusionPipeline.from_pretrained(
-    "gsdf/Counterfeit-V2.5", torch_dtype=torch.float16, safety_checker=None
+    "gsdf/Counterfeit-V2.5", torch_dtype=torch.float16, safety_checker=None, use_safetensors=True
 ).to("cuda")
 pipeline.scheduler = DPMSolverMultistepScheduler.from_config(
     pipeline.scheduler.config, use_karras_sigmas=True

docs/source/en/training/text2image.md

@@ -238,7 +238,7 @@ Now you can load the fine-tuned model for inference by passing the model path or
 from diffusers import StableDiffusionPipeline
 
 model_path = "path_to_saved_model"
-pipe = StableDiffusionPipeline.from_pretrained(model_path, torch_dtype=torch.float16)
+pipe = StableDiffusionPipeline.from_pretrained(model_path, torch_dtype=torch.float16, use_safetensors=True)
 pipe.to("cuda")
 
 image = pipe(prompt="yoda").images[0]

docs/source/en/training/text_inversion.md

@@ -204,7 +204,7 @@ from diffusers import StableDiffusionPipeline
 import torch
 
 model_id = "runwayml/stable-diffusion-v1-5"
-pipe = StableDiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.float16).to("cuda")
+pipe = StableDiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.float16, use_safetensors=True).to("cuda")
 ```
 
 Next, we need to load the textual inversion embedding vector which can be done via the [`TextualInversionLoaderMixin.load_textual_inversion`]

docs/source/en/using-diffusers/conditional_image_generation.md

@@ -25,7 +25,7 @@ In this guide, you'll use [`DiffusionPipeline`] for text-to-image generation wit
 ```python
 >>> from diffusers import DiffusionPipeline
 
->>> generator = DiffusionPipeline.from_pretrained("runwayml/stable-diffusion-v1-5")
+>>> generator = DiffusionPipeline.from_pretrained("runwayml/stable-diffusion-v1-5", use_safetensors=True)
 ```
 
 The [`DiffusionPipeline`] downloads and caches all modeling, tokenization, and scheduling components. 

docs/source/en/using-diffusers/contribute_pipeline.md

@@ -94,7 +94,7 @@ output = pipeline()
 But what's even better is you can load pre-existing weights into the pipeline if the pipeline structure is identical. For example, you can load the [`google/ddpm-cifar10-32`](https://huggingface.co/google/ddpm-cifar10-32) weights into the one-step pipeline:
 
 ```python
-pipeline = UnetSchedulerOneForwardPipeline.from_pretrained("google/ddpm-cifar10-32")
+pipeline = UnetSchedulerOneForwardPipeline.from_pretrained("google/ddpm-cifar10-32", use_safetensors=True)
 
 output = pipeline()
 ```
@@ -108,7 +108,9 @@ Once it is merged, anyone with `diffusers >= 0.4.0` installed can use this pipel
 ```python
 from diffusers import DiffusionPipeline
 
-pipe = DiffusionPipeline.from_pretrained("google/ddpm-cifar10-32", custom_pipeline="one_step_unet")
+pipe = DiffusionPipeline.from_pretrained(
+    "google/ddpm-cifar10-32", custom_pipeline="one_step_unet", use_safetensors=True
+)
 pipe()
 ```
 
@@ -117,7 +119,9 @@ Another way to share your community pipeline is to upload the `one_step_unet.py`
 ```python
 from diffusers import DiffusionPipeline
 
-pipeline = DiffusionPipeline.from_pretrained("google/ddpm-cifar10-32", custom_pipeline="stevhliu/one_step_unet")
+pipeline = DiffusionPipeline.from_pretrained(
+    "google/ddpm-cifar10-32", custom_pipeline="stevhliu/one_step_unet", use_safetensors=True
+)
 ```
 
 Take a look at the following table to compare the two sharing workflows to help you decide the best option for you:
@@ -161,6 +165,7 @@ pipeline = DiffusionPipeline.from_pretrained(
     feature_extractor=feature_extractor,
     scheduler=scheduler,
     torch_dtype=torch.float16,
+    use_safetensors=True,
 )
 ```
 

docs/source/en/using-diffusers/control_brightness.md

@@ -24,7 +24,7 @@ Next, configure the following parameters in the [`DDIMScheduler`]:
 ```py
 >>> from diffusers import DiffusionPipeline, DDIMScheduler
 
->>> pipeline = DiffusionPipeline.from_pretrained("ptx0/pseudo-journey-v2")
+>>> pipeline = DiffusionPipeline.from_pretrained("ptx0/pseudo-journey-v2", use_safetensors=True)
 # switch the scheduler in the pipeline to use the DDIMScheduler
 
 >>> pipeline.scheduler = DDIMScheduler.from_config(

docs/source/en/using-diffusers/custom_pipeline_examples.md

@@ -32,7 +32,7 @@ If a community doesn't work as expected, please open an issue and ping the autho
 To load a custom pipeline you just need to pass the `custom_pipeline` argument to `DiffusionPipeline`, as one of the files in `diffusers/examples/community`. Feel free to send a PR with your own pipelines, we will merge them quickly.
 ```py
 pipe = DiffusionPipeline.from_pretrained(
-    "CompVis/stable-diffusion-v1-4", custom_pipeline="filename_in_the_community_folder"
+    "CompVis/stable-diffusion-v1-4", custom_pipeline="filename_in_the_community_folder", use_safetensors=True
 )
 ```
 
@@ -61,6 +61,7 @@ guided_pipeline = DiffusionPipeline.from_pretrained(
     clip_model=clip_model,
     feature_extractor=feature_extractor,
     torch_dtype=torch.float16,
+    use_safetensors=True,
 )
 guided_pipeline.enable_attention_slicing()
 guided_pipeline = guided_pipeline.to("cuda")
@@ -117,6 +118,7 @@ pipe = DiffusionPipeline.from_pretrained(
     torch_dtype=torch.float16,
     safety_checker=None,  # Very important for videos...lots of false positives while interpolating
     custom_pipeline="interpolate_stable_diffusion",
+    use_safetensors=True,
 ).to("cuda")
 pipe.enable_attention_slicing()
 
@@ -159,6 +161,7 @@ pipe = DiffusionPipeline.from_pretrained(
     "CompVis/stable-diffusion-v1-4",
     custom_pipeline="stable_diffusion_mega",
     torch_dtype=torch.float16,
+    use_safetensors=True,
 )
 pipe.to("cuda")
 pipe.enable_attention_slicing()
@@ -203,7 +206,7 @@ from diffusers import DiffusionPipeline
 import torch
 
 pipe = DiffusionPipeline.from_pretrained(
-    "hakurei/waifu-diffusion", custom_pipeline="lpw_stable_diffusion", torch_dtype=torch.float16
+    "hakurei/waifu-diffusion", custom_pipeline="lpw_stable_diffusion", torch_dtype=torch.float16, use_safetensors=True
 )
 pipe = pipe.to("cuda")
 
@@ -224,6 +227,7 @@ pipe = DiffusionPipeline.from_pretrained(
     custom_pipeline="lpw_stable_diffusion_onnx",
     revision="onnx",
     provider="CUDAExecutionProvider",
+    use_safetensors=True,
 )
 
 prompt = "a photo of an astronaut riding a horse on mars, best quality"
@@ -267,8 +271,8 @@ diffuser_pipeline = DiffusionPipeline.from_pretrained(
     custom_pipeline="speech_to_image_diffusion",
     speech_model=model,
     speech_processor=processor,
-    
     torch_dtype=torch.float16,
+    use_safetensors=True,
 )
 
 diffuser_pipeline.enable_attention_slicing()

docs/source/en/using-diffusers/custom_pipeline_overview.md

@@ -30,7 +30,7 @@ To load any community pipeline on the Hub, pass the repository id of the communi
 from diffusers import DiffusionPipeline
 
 pipeline = DiffusionPipeline.from_pretrained(
-    "google/ddpm-cifar10-32", custom_pipeline="hf-internal-testing/diffusers-dummy-pipeline"
+    "google/ddpm-cifar10-32", custom_pipeline="hf-internal-testing/diffusers-dummy-pipeline", use_safetensors=True
 )
 ```
 
@@ -50,6 +50,7 @@ pipeline = DiffusionPipeline.from_pretrained(
     custom_pipeline="clip_guided_stable_diffusion",
     clip_model=clip_model,
     feature_extractor=feature_extractor,
+    use_safetensors=True,
 )
 ```
 

docs/source/en/using-diffusers/depth2img.md

@@ -28,6 +28,7 @@ from diffusers import StableDiffusionDepth2ImgPipeline
 pipe = StableDiffusionDepth2ImgPipeline.from_pretrained(
     "stabilityai/stable-diffusion-2-depth",
     torch_dtype=torch.float16,
+    use_safetensors=True,
 ).to("cuda")
 ```
 

docs/source/en/using-diffusers/img2img.md

@@ -33,9 +33,9 @@ from io import BytesIO
 from diffusers import StableDiffusionImg2ImgPipeline
 
 device = "cuda"
-pipe = StableDiffusionImg2ImgPipeline.from_pretrained("nitrosocke/Ghibli-Diffusion", torch_dtype=torch.float16).to(
-    device
-)
+pipe = StableDiffusionImg2ImgPipeline.from_pretrained(
+    "nitrosocke/Ghibli-Diffusion", torch_dtype=torch.float16, use_safetensors=True
+).to(device)
 ```
 
 Download and preprocess an initial image so you can pass it to the pipeline:

docs/source/en/using-diffusers/inpaint.md

@@ -29,6 +29,7 @@ from diffusers import StableDiffusionInpaintPipeline
 pipeline = StableDiffusionInpaintPipeline.from_pretrained(
     "runwayml/stable-diffusion-inpainting",
     torch_dtype=torch.float16,
+    use_safetensors=True,
 )
 pipeline = pipeline.to("cuda")
 ```

docs/source/en/using-diffusers/loading.md

@@ -39,7 +39,7 @@ The [`DiffusionPipeline`] class is the simplest and most generic way to load any
 from diffusers import DiffusionPipeline
 
 repo_id = "runwayml/stable-diffusion-v1-5"
-pipe = DiffusionPipeline.from_pretrained(repo_id)
+pipe = DiffusionPipeline.from_pretrained(repo_id, use_safetensors=True)
 ```
 
 You can also load a checkpoint with it's specific pipeline class. The example above loaded a Stable Diffusion model; to get the same result, use the [`StableDiffusionPipeline`] class:
@@ -48,7 +48,7 @@ You can also load a checkpoint with it's specific pipeline class. The example ab
 from diffusers import StableDiffusionPipeline
 
 repo_id = "runwayml/stable-diffusion-v1-5"
-pipe = StableDiffusionPipeline.from_pretrained(repo_id)
+pipe = StableDiffusionPipeline.from_pretrained(repo_id, use_safetensors=True)
 ```
 
 A checkpoint (such as [`CompVis/stable-diffusion-v1-4`](https://huggingface.co/CompVis/stable-diffusion-v1-4) or [`runwayml/stable-diffusion-v1-5`](https://huggingface.co/runwayml/stable-diffusion-v1-5)) may also be used for more than one task, like text-to-image or image-to-image. To differentiate what task you want to use the checkpoint for, you have to load it directly with it's corresponding task-specific pipeline class:
@@ -75,7 +75,7 @@ Then pass the local path to [`~DiffusionPipeline.from_pretrained`]:
 from diffusers import DiffusionPipeline
 
 repo_id = "./stable-diffusion-v1-5"
-stable_diffusion = DiffusionPipeline.from_pretrained(repo_id)
+stable_diffusion = DiffusionPipeline.from_pretrained(repo_id, use_safetensors=True)
 ```
 
 The [`~DiffusionPipeline.from_pretrained`] method won't download any files from the Hub when it detects a local path, but this also means it won't download and cache the latest changes to a checkpoint.
@@ -94,7 +94,7 @@ To find out which schedulers are compatible for customization, you can use the `
 from diffusers import DiffusionPipeline
 
 repo_id = "runwayml/stable-diffusion-v1-5"
-stable_diffusion = DiffusionPipeline.from_pretrained(repo_id)
+stable_diffusion = DiffusionPipeline.from_pretrained(repo_id, use_safetensors=True)
 stable_diffusion.scheduler.compatibles
 ```
 
@@ -109,7 +109,7 @@ repo_id = "runwayml/stable-diffusion-v1-5"
 
 scheduler = EulerDiscreteScheduler.from_pretrained(repo_id, subfolder="scheduler")
 
-stable_diffusion = DiffusionPipeline.from_pretrained(repo_id, scheduler=scheduler)
+stable_diffusion = DiffusionPipeline.from_pretrained(repo_id, scheduler=scheduler, use_safetensors=True)
 ```
 
 ### Safety checker
@@ -120,7 +120,7 @@ Diffusion models like Stable Diffusion can generate harmful content, which is wh
 from diffusers import DiffusionPipeline
 
 repo_id = "runwayml/stable-diffusion-v1-5"
-stable_diffusion = DiffusionPipeline.from_pretrained(repo_id, safety_checker=None)
+stable_diffusion = DiffusionPipeline.from_pretrained(repo_id, safety_checker=None, use_safetensors=True)
 ```
 
 ### Reuse components across pipelines
@@ -131,7 +131,7 @@ You can also reuse the same components in multiple pipelines to avoid loading th
 from diffusers import StableDiffusionPipeline, StableDiffusionImg2ImgPipeline
 
 model_id = "runwayml/stable-diffusion-v1-5"
-stable_diffusion_txt2img = StableDiffusionPipeline.from_pretrained(model_id)
+stable_diffusion_txt2img = StableDiffusionPipeline.from_pretrained(model_id, use_safetensors=True)
 
 components = stable_diffusion_txt2img.components
 ```
@@ -148,7 +148,7 @@ You can also pass the components individually to the pipeline if you want more f
 from diffusers import StableDiffusionPipeline, StableDiffusionImg2ImgPipeline
 
 model_id = "runwayml/stable-diffusion-v1-5"
-stable_diffusion_txt2img = StableDiffusionPipeline.from_pretrained(model_id)
+stable_diffusion_txt2img = StableDiffusionPipeline.from_pretrained(model_id, use_safetensors=True)
 stable_diffusion_img2img = StableDiffusionImg2ImgPipeline(
     vae=stable_diffusion_txt2img.vae,
     text_encoder=stable_diffusion_txt2img.text_encoder,
@@ -194,10 +194,12 @@ import torch
 
 # load fp16 variant
 stable_diffusion = DiffusionPipeline.from_pretrained(
-    "runwayml/stable-diffusion-v1-5", variant="fp16", torch_dtype=torch.float16
+    "runwayml/stable-diffusion-v1-5", variant="fp16", torch_dtype=torch.float16, use_safetensors=True
 )
 # load non_ema variant
-stable_diffusion = DiffusionPipeline.from_pretrained("runwayml/stable-diffusion-v1-5", variant="non_ema")
+stable_diffusion = DiffusionPipeline.from_pretrained(
+    "runwayml/stable-diffusion-v1-5", variant="non_ema", use_safetensors=True
+)
 ```
 
 To save a checkpoint stored in a different floating point type or as a non-EMA variant, use the [`DiffusionPipeline.save_pretrained`] method and specify the `variant` argument. You should try and save a variant to the same folder as the original checkpoint, so you can load both from the same folder:
@@ -215,10 +217,12 @@ If you don't save the variant to an existing folder, you must specify the `varia
 
 ```python
 # 👎 this won't work
-stable_diffusion = DiffusionPipeline.from_pretrained("./stable-diffusion-v1-5", torch_dtype=torch.float16)
+stable_diffusion = DiffusionPipeline.from_pretrained(
+    "./stable-diffusion-v1-5", torch_dtype=torch.float16, use_safetensors=True
+)
 # 👍 this works
 stable_diffusion = DiffusionPipeline.from_pretrained(
-    "./stable-diffusion-v1-5", variant="fp16", torch_dtype=torch.float16
+    "./stable-diffusion-v1-5", variant="fp16", torch_dtype=torch.float16, use_safetensors=True
 )
 ```
 
@@ -233,7 +237,7 @@ load model variants, e.g.:
 ```python
 from diffusers import DiffusionPipeline
 
-pipe = DiffusionPipeline.from_pretrained("CompVis/stable-diffusion-v1-4", revision="fp16")
+pipe = DiffusionPipeline.from_pretrained("CompVis/stable-diffusion-v1-4", revision="fp16", use_safetensors=True)
 ```
 
 However, this behavior is now deprecated since the "revision" argument should (just as it's done in GitHub) better be used to load model checkpoints from a specific commit or branch in development.
@@ -259,7 +263,7 @@ Models can be loaded from a subfolder with the `subfolder` argument. For example
 from diffusers import UNet2DConditionModel
 
 repo_id = "runwayml/stable-diffusion-v1-5"
-model = UNet2DConditionModel.from_pretrained(repo_id, subfolder="unet")
+model = UNet2DConditionModel.from_pretrained(repo_id, subfolder="unet", use_safetensors=True)
 ```
 
 Or directly from a repository's [directory](https://huggingface.co/google/ddpm-cifar10-32/tree/main):
@@ -268,7 +272,7 @@ Or directly from a repository's [directory](https://huggingface.co/google/ddpm-c
 from diffusers import UNet2DModel
 
 repo_id = "google/ddpm-cifar10-32"
-model = UNet2DModel.from_pretrained(repo_id)
+model = UNet2DModel.from_pretrained(repo_id, use_safetensors=True)
 ```
 
 You can also load and save model variants by specifying the `variant` argument in [`ModelMixin.from_pretrained`] and [`ModelMixin.save_pretrained`]:
@@ -276,7 +280,9 @@ You can also load and save model variants by specifying the `variant` argument i
 ```python
 from diffusers import UNet2DConditionModel
 
-model = UNet2DConditionModel.from_pretrained("runwayml/stable-diffusion-v1-5", subfolder="unet", variant="non-ema")
+model = UNet2DConditionModel.from_pretrained(
+    "runwayml/stable-diffusion-v1-5", subfolder="unet", variant="non-ema", use_safetensors=True
+)
 model.save_pretrained("./local-unet", variant="non-ema")
 ```
 
@@ -310,7 +316,7 @@ euler = EulerDiscreteScheduler.from_pretrained(repo_id, subfolder="scheduler")
 dpm = DPMSolverMultistepScheduler.from_pretrained(repo_id, subfolder="scheduler")
 
 # replace `dpm` with any of `ddpm`, `ddim`, `pndm`, `lms`, `euler_anc`, `euler`
-pipeline = StableDiffusionPipeline.from_pretrained(repo_id, scheduler=dpm)
+pipeline = StableDiffusionPipeline.from_pretrained(repo_id, scheduler=dpm, use_safetensors=True)
 ```
 
 ## DiffusionPipeline explained
@@ -326,7 +332,7 @@ The pipelines underlying folder structure corresponds directly with their class
 from diffusers import DiffusionPipeline
 
 repo_id = "runwayml/stable-diffusion-v1-5"
-pipeline = DiffusionPipeline.from_pretrained(repo_id)
+pipeline = DiffusionPipeline.from_pretrained(repo_id, use_safetensors=True)
 print(pipeline)
 ```
 

docs/source/en/using-diffusers/other-formats.md

@@ -111,7 +111,9 @@ If you prefer to run inference with code, click on the **Use in Diffusers** butt
 ```py
 from diffusers import DiffusionPipeline
 
-pipeline = DiffusionPipeline.from_pretrained("sayakpaul/textual-inversion-cat-kerascv_sd_diffusers_pipeline")
+pipeline = DiffusionPipeline.from_pretrained(
+    "sayakpaul/textual-inversion-cat-kerascv_sd_diffusers_pipeline", use_safetensors=True
+)
 ```
 
 Then you can generate an image like:
@@ -119,7 +121,9 @@ Then you can generate an image like:
 ```py
 from diffusers import DiffusionPipeline
 
-pipeline = DiffusionPipeline.from_pretrained("sayakpaul/textual-inversion-cat-kerascv_sd_diffusers_pipeline")
+pipeline = DiffusionPipeline.from_pretrained(
+    "sayakpaul/textual-inversion-cat-kerascv_sd_diffusers_pipeline", use_safetensors=True
+)
 pipeline.to("cuda")
 
 placeholder_token = "<my-funny-cat-token>"

docs/source/en/using-diffusers/reproducibility.md

@@ -40,7 +40,7 @@ import numpy as np
 model_id = "google/ddpm-cifar10-32"
 
 # load model and scheduler
-ddim = DDIMPipeline.from_pretrained(model_id)
+ddim = DDIMPipeline.from_pretrained(model_id, use_safetensors=True)
 
 # run pipeline for just two steps and return numpy tensor
 image = ddim(num_inference_steps=2, output_type="np").images
@@ -65,7 +65,7 @@ import numpy as np
 model_id = "google/ddpm-cifar10-32"
 
 # load model and scheduler
-ddim = DDIMPipeline.from_pretrained(model_id)
+ddim = DDIMPipeline.from_pretrained(model_id, use_safetensors=True)
 
 # create a generator for reproducibility
 generator = torch.Generator(device="cpu").manual_seed(0)
@@ -100,7 +100,7 @@ import numpy as np
 model_id = "google/ddpm-cifar10-32"
 
 # load model and scheduler
-ddim = DDIMPipeline.from_pretrained(model_id)
+ddim = DDIMPipeline.from_pretrained(model_id, use_safetensors=True)
 ddim.to("cuda")
 
 # create a generator for reproducibility
@@ -125,7 +125,7 @@ import numpy as np
 model_id = "google/ddpm-cifar10-32"
 
 # load model and scheduler
-ddim = DDIMPipeline.from_pretrained(model_id)
+ddim = DDIMPipeline.from_pretrained(model_id, use_safetensors=True)
 ddim.to("cuda")
 
 # create a generator for reproducibility; notice you don't place it on the GPU!
@@ -174,7 +174,7 @@ from diffusers import DDIMScheduler, StableDiffusionPipeline
 import numpy as np
 
 model_id = "runwayml/stable-diffusion-v1-5"
-pipe = StableDiffusionPipeline.from_pretrained(model_id).to("cuda")
+pipe = StableDiffusionPipeline.from_pretrained(model_id, use_safetensors=True).to("cuda")
 pipe.scheduler = DDIMScheduler.from_config(pipe.scheduler.config)
 g = torch.Generator(device="cuda")
 

docs/source/en/using-diffusers/reusing_seeds.md

@@ -27,7 +27,9 @@ Instantiate a pipeline with [`DiffusionPipeline.from_pretrained`] and place it o
 ```python
 >>> from diffusers import DiffusionPipeline
 
->>> pipe = DiffusionPipeline.from_pretrained("runwayml/stable-diffusion-v1-5", torch_dtype=torch.float16)
+>>> pipe = DiffusionPipeline.from_pretrained(
+...     "runwayml/stable-diffusion-v1-5", torch_dtype=torch.float16, use_safetensors=True
+... )
 >>> pipe = pipe.to("cuda")
 ```
 

docs/source/en/using-diffusers/schedulers.md

@@ -39,7 +39,9 @@ import torch
 
 login()
 
-pipeline = DiffusionPipeline.from_pretrained("runwayml/stable-diffusion-v1-5", torch_dtype=torch.float16)
+pipeline = DiffusionPipeline.from_pretrained(
+    "runwayml/stable-diffusion-v1-5", torch_dtype=torch.float16, use_safetensors=True
+)
 ```
 
 Next, we move it to GPU:

docs/source/en/using-diffusers/textual_inversion_inference.md

@@ -49,7 +49,9 @@ repo_id_embeds = "sd-concepts-library/cat-toy"
 Now you can load a pipeline, and pass the pre-learned concept to it:
 
 ```py
-pipeline = StableDiffusionPipeline.from_pretrained(pretrained_model_name_or_path, torch_dtype=torch.float16).to("cuda")
+pipeline = StableDiffusionPipeline.from_pretrained(
+    pretrained_model_name_or_path, torch_dtype=torch.float16, use_safetensors=True
+).to("cuda")
 
 pipeline.load_textual_inversion(repo_id_embeds)
 ```

docs/source/en/using-diffusers/unconditional_image_generation.md

@@ -32,7 +32,7 @@ In this guide, you'll use [`DiffusionPipeline`] for unconditional image generati
 ```python
 >>> from diffusers import DiffusionPipeline
 
->>> generator = DiffusionPipeline.from_pretrained("anton-l/ddpm-butterflies-128")
+>>> generator = DiffusionPipeline.from_pretrained("anton-l/ddpm-butterflies-128", use_safetensors=True)
 ```
 
 The [`DiffusionPipeline`] downloads and caches all modeling, tokenization, and scheduling components. 

docs/source/en/using-diffusers/using_safetensors.md

@@ -40,7 +40,9 @@ You can use the model with the new `.safetensors` weights by specifying the refe
 ```py
 from diffusers import DiffusionPipeline
 
-pipeline = DiffusionPipeline.from_pretrained("stabilityai/stable-diffusion-2-1", revision="refs/pr/22")
+pipeline = DiffusionPipeline.from_pretrained(
+    "stabilityai/stable-diffusion-2-1", revision="refs/pr/22", use_safetensors=True
+)
 ```
 
 ## Why use safetensors?
@@ -55,7 +57,7 @@ There are several reasons for using safetensors:
 	```py
  from diffusers import StableDiffusionPipeline
 
- pipeline = StableDiffusionPipeline.from_pretrained("stabilityai/stable-diffusion-2-1")
+ pipeline = StableDiffusionPipeline.from_pretrained("stabilityai/stable-diffusion-2-1", use_safetensors=True)
  "Loaded in safetensors 0:00:02.033658"
  "Loaded in PyTorch 0:00:02.663379"
 	```

docs/source/en/using-diffusers/write_own_pipeline.md

@@ -25,7 +25,7 @@ A pipeline is a quick and easy way to run a model for inference, requiring no mo
 ```py
 >>> from diffusers import DDPMPipeline
 
->>> ddpm = DDPMPipeline.from_pretrained("google/ddpm-cat-256").to("cuda")
+>>> ddpm = DDPMPipeline.from_pretrained("google/ddpm-cat-256", use_safetensors=True).to("cuda")
 >>> image = ddpm(num_inference_steps=25).images[0]
 >>> image
 ```
@@ -46,7 +46,7 @@ To recreate the pipeline with the model and scheduler separately, let's write ou
 >>> from diffusers import DDPMScheduler, UNet2DModel
 
 >>> scheduler = DDPMScheduler.from_pretrained("google/ddpm-cat-256")
->>> model = UNet2DModel.from_pretrained("google/ddpm-cat-256").to("cuda")
+>>> model = UNet2DModel.from_pretrained("google/ddpm-cat-256", use_safetensors=True).to("cuda")
 ```
 
 2. Set the number of timesteps to run the denoising process for:
@@ -124,10 +124,14 @@ Now that you know what you need for the Stable Diffusion pipeline, load all thes
 >>> from transformers import CLIPTextModel, CLIPTokenizer
 >>> from diffusers import AutoencoderKL, UNet2DConditionModel, PNDMScheduler
 
->>> vae = AutoencoderKL.from_pretrained("CompVis/stable-diffusion-v1-4", subfolder="vae")
+>>> vae = AutoencoderKL.from_pretrained("CompVis/stable-diffusion-v1-4", subfolder="vae", use_safetensors=True)
 >>> tokenizer = CLIPTokenizer.from_pretrained("CompVis/stable-diffusion-v1-4", subfolder="tokenizer")
->>> text_encoder = CLIPTextModel.from_pretrained("CompVis/stable-diffusion-v1-4", subfolder="text_encoder")
->>> unet = UNet2DConditionModel.from_pretrained("CompVis/stable-diffusion-v1-4", subfolder="unet")
+>>> text_encoder = CLIPTextModel.from_pretrained(
+...     "CompVis/stable-diffusion-v1-4", subfolder="text_encoder", use_safetensors=True
+... )
+>>> unet = UNet2DConditionModel.from_pretrained(
+...     "CompVis/stable-diffusion-v1-4", subfolder="unet", use_safetensors=True
+... )
 ```
 
 Instead of the default [`PNDMScheduler`], exchange it for the [`UniPCMultistepScheduler`] to see how easy it is to plug a different scheduler in: